Majority-Based Tracking Forecast Memories for Stochastic LDPC Decoding
Saeed Sharifi Tehrani, Ali Naderi, Guy-Armand Kamendje, Saied Hemati, Shie Mannor, Warren J. Gross · IEEE Transactions on Signal Processing · 2010
This paper proposes majority-based tracking forecast memories (MTFMs) for area efficient high throughput ASIC implementation of stochastic Low-Density Parity-Check (LDPC) decoders. The proposed method is applied for ASIC implementation of a fully parallel stochastic decoder that decodes the (2048, 1723) LDPC code from the IEEE 802.3an (10GBASE-T) standard. The decoder occupies a silicon core area of 6.38${\hbox {mm}}^{2}$in CMOS 90 nm technology, achieves a maximum clock frequency of 500 MHz, and provides a maximum core throughput of 61.3 Gb/s. The decoder also has good decoding performance and error-floor behavior and provides a bit error rate (BER) of about$4\times 10^{-13}$at$E_{b}/N_{0}={\hbox {5.15\ dB}}$. To the best of our knowledge, the implemented decoder is the most area efficient fully parallel soft -decision LDPC decoder reported in the literature.